Problem
Can we predict rain fall and bicycle traffic using atmospheric data from the attic of a Cambridge apartment?
Bike study
Estimate bicycle traffic based on the temperature, humidity, light, and pressure readings of an attic-mounted sensor.
Rain study
Classify rainfall based on temperature, humidity, light, and pressure in an attic-mounted sensor.
Data Collection
To begin collecting data, we installed two sensors in a Cambridge apartment. One went upstairs, in the attic, while the other was in the living room. We encountered many issues with our two sensors.
Reset issue
The built-in soft reset malfunctioned, requiring us to perform a hard reset before each collection.
sd card issue
The SD card reader shifted in the sensor, causing improper insertions on two occasions.
Duration issue
Setting the collection time to 99 hours prevented a full save of our data, ruining several collection periods.
broken sensor
We initially planned to use data from two sensors, but one of our sensors broke in early November.
Ultimately, we were only able to collect six days of data on one sensor, rather than the full month of data from two sensors that we had hoped for. At this point, we decided to bring in some external datasets to augment our project, using two publicly available datasets.
RAIN DATA
BIKE DATA
One of our datasets measures Boston rainfall. The other measures Cambridge bicycle traffic. Click on the links at left to see more.
Exploratory Analysis
Once we had small samples of data, we performed some exploratory analysis to verify that we could indeed attempt to estimate bicycle traffic using the attic sensors.
Clockwise from top left: our sensor, the attic where we gathered data, code to read in the data, a summary of the initial data.
Visit the Findings pages to see what we learned after collecting our data!